even2004： CFD - Divergence of the linear iterations错误的解
CFD - Divergence of the linear iterations错误的解决办法
hi!
I had the same problem. If you take a look at the examples 'Integrated square-shaped inductor' (model library -> mulitphysic ->electromagnetics ) and 'power conductor' (model library -> Ac/Dc module -> electrical components) you can see, that in theses examples the boundaries on which the current /potential is applied is in the same level as the boundarie, that sourrounds the whole geometrie.
I tried it with my probelm, and now it works perfectly!
do not ask how long it took me to find this solution ^^
best wishes!
michael
--- In  COMSOL_Users@yahoogroups.com, "sfwei01" wrote:
>
> I also met such similar problems when dealing with 3D electromagnetic model. So far, I still haven't found a better solution.
>
> --- In COMSOL_Users@yahoogroups.com, "ralfhinas" wrote:
> >
> > Hi everybody!
> >
> > I'm quite familiar with CFD and COMSOL. And get the message shown below
> > VERY OFTEN
> >
> > Error:
> > Failed to find a solution:
> > Divergence of the linear iterations.
> > Returned solution has not converged.
> >
> > You can sometimes avoid it by setting the right parameters, but when
> > using a similar model, everything changes and you must restart finding
> > the appropriate parameters which makes comparing models really
> > difficult.
> >
> > I've read almost the whole literature and postings but still don't have
> > a
> > how-to rule applicable and reliable to all (and not only some
> > simplified) 3D!-geometries.
> >
> > Settings are usually: P-in, P-out bla bla, NS-incomp bla bla, flow rate
> > bla bla bla(
> >
> > Does anyone has THE hint for me how to proceed?
> >
> > The comsol support gave me the only advice to set a higher value for the
> > Newton iteration number, but that can't be the problem anymore!
> >
> > Would be lad if anyone has the rule applicable to a wide range of models!
> > Thanks for your assistance!!!!
> >
> > Antonius in Berlin
|
|
阳光精品论坛 >> 『 CFD 专栏 』 >> 回复文章